Five common pitfalls in eye-tracking research

Abstract

Eye tracking has become an increasingly popular research methodology among language researchers to examine online cognitive processing of second-language (L2) speakers and bilinguals. As the scope of eye-tracking research expands, there is a need to ensure that the methodology is used properly, so as to safeguard the validity of research findings and the empirical basis upon which theories are built. We present five pitfalls in eye-tracking research that may threaten the internal or external validity of the study. We offer concrete recommendations to avoid each pitfall and demonstrate how these recommendations can be implemented so researchers may take full advantage of the opportunities of eye-tracking research. The five pitfalls occur at all stages of the research cycle. We emphasize (1) adopting a confirmatory (theory-driven, hypothesis-testing) approach to research, (2) selecting and handling eye-movement measures in a principled manner, (3) aligning the research design with the eye tracker properties (spatial accuracy, precision), (4) inspecting raw, sample-level eye-tracking data and their preprocessing by the eye-tracking software. We conclude by inviting eye-tracking researchers (5) to embrace open science practices to enhance the replication and reproducibility of their work.